57369
|
1 |
/*
|
|
2 |
* Copyright (c) 2016, 2019, Oracle and/or its affiliates. All rights reserved.
|
|
3 |
* ORACLE PROPRIETARY/CONFIDENTIAL. Use is subject to license terms.
|
|
4 |
*
|
|
5 |
*
|
|
6 |
*
|
|
7 |
*
|
|
8 |
*
|
|
9 |
*
|
|
10 |
*
|
|
11 |
*
|
|
12 |
*
|
|
13 |
*
|
|
14 |
*
|
|
15 |
*
|
|
16 |
*
|
|
17 |
*
|
|
18 |
*
|
|
19 |
*
|
|
20 |
*
|
|
21 |
*
|
|
22 |
*
|
|
23 |
*
|
|
24 |
*/
|
|
25 |
|
|
26 |
// package java.util;
|
|
27 |
|
|
28 |
import java.math.BigInteger;
|
|
29 |
import java.util.concurrent.atomic.AtomicLong;
|
|
30 |
|
|
31 |
/**
|
|
32 |
* A generator of uniform pseudorandom values applicable for use in
|
|
33 |
* (among other contexts) isolated parallel computations that may
|
|
34 |
* generate subtasks. Class {@code L64X1024Random} implements
|
|
35 |
* interfaces {@link java.util.Rng} and {@link java.util.SplittableRng},
|
|
36 |
* and therefore supports methods for producing pseudorandomly chosen
|
|
37 |
* numbers of type {@code int}, {@code long}, {@code float}, and {@code double}
|
|
38 |
* as well as creating new split-off {@code L64X1024Random} objects,
|
|
39 |
* with similar usages as for class {@link java.util.SplittableRandom}.
|
|
40 |
*
|
|
41 |
* <p>Series of generated values pass the TestU01 BigCrush and PractRand test suites
|
|
42 |
* that measure independence and uniformity properties of random number generators.
|
|
43 |
* (Most recently validated with
|
|
44 |
* <a href="http://simul.iro.umontreal.ca/testu01/tu01.html">version 1.2.3 of TestU01</a>
|
|
45 |
* and <a href="http://pracrand.sourceforge.net">version 0.90 of PractRand</a>.
|
|
46 |
* Note that TestU01 BigCrush was used to test not only values produced by the {@code nextLong()}
|
|
47 |
* method but also the result of bit-reversing each value produced by {@code nextLong()}.)
|
|
48 |
* These tests validate only the methods for certain
|
|
49 |
* types and ranges, but similar properties are expected to hold, at
|
|
50 |
* least approximately, for others as well.
|
|
51 |
*
|
|
52 |
* <p>{@code L64X1024Random} is a specific member of the LXM family of algorithms
|
|
53 |
* for pseudorandom number generators. Every LXM generator consists of two
|
|
54 |
* subgenerators; one is an LCG (Linear Congruential Generator) and the other is
|
|
55 |
* an Xorshift generator. Each output of an LXM generator is the sum of one
|
|
56 |
* output from each subgenerator, possibly processed by a final mixing function
|
|
57 |
* (but {@code L64X1024Random} does not use a mixing function).
|
|
58 |
*
|
|
59 |
* <p>The LCG subgenerator for {@code L64X1024Random} has an update step of the
|
|
60 |
* form {@code s = m * s + a}, where {@code s}, {@code m}, and {@code a} are all
|
|
61 |
* of type {@code long}; {@code s} is the mutable state, the multiplier {@code m}
|
|
62 |
* is fixed (the same for all instances of {@code L64X1024Random}}) and the addend
|
|
63 |
* {@code a} is a parameter (a final field of the instance). The parameter
|
|
64 |
* {@code a} is required to be odd (this allows the LCG to have the maximal
|
|
65 |
* period, namely 2<sup>64</sup>); therefore there are 2<sup>63</sup> distinct choices
|
|
66 |
* of parameter.
|
|
67 |
*
|
|
68 |
* <p>The Xorshift subgenerator for {@code L64X1024Random} is the {@code xoroshiro1024}
|
|
69 |
* algorithm (parameters 25, 27, and 36), without any final scrambler such as "+" or "**".
|
|
70 |
* Its state consists of an array {@code x} of sixteen {@code long} values,
|
|
71 |
* which can take on any values provided that they are not all zero.
|
|
72 |
* The period of this subgenerator is 2<sup>1024</sup>-1.
|
|
73 |
*
|
|
74 |
* <p> Because the periods 2<sup>64</sup> and 2<sup>1024</sup>-1 of the two subgenerators
|
|
75 |
* are relatively prime, the <em>period</em> of any single {@code L64X1024Random} object
|
|
76 |
* (the length of the series of generated 64-bit values before it repeats) is the product
|
|
77 |
* of the periods of the subgenerators, that is, 2<sup>64</sup>(2<sup>1024</sup>-1),
|
|
78 |
* which is just slightly smaller than 2<sup>1088</sup>. Moreover, if two distinct
|
|
79 |
* {@code L64X1024Random} objects have different {@code a} parameters, then their
|
|
80 |
* cycles of produced values will be different.
|
|
81 |
*
|
|
82 |
* <p>The 64-bit values produced by the {@code nextLong()} method are exactly equidistributed.
|
|
83 |
* For any specific instance of {@code L64X1024Random}, over the course of its cycle each
|
|
84 |
* of the 2<sup>64</sup> possible {@code long} values will be produced 2<sup>1024</sup>-1 times.
|
|
85 |
* The values produced by the {@code nextInt()}, {@code nextFloat()}, and {@code nextDouble()}
|
|
86 |
* methods are likewise exactly equidistributed.
|
|
87 |
*
|
|
88 |
* <p>In fact, the 64-bit values produced by the {@code nextLong()} method are 16-equidistributed.
|
|
89 |
* To be precise: for any specific instance of {@code L64X1024Random}, consider
|
|
90 |
* the (overlapping) length-16 subsequences of the cycle of 64-bit values produced by
|
|
91 |
* {@code nextLong()} (assuming no other methods are called that would affect the state).
|
|
92 |
* There are 2<sup>64</sup>(2<sup>1024</sup>-1) such subsequences, and each subsequence,
|
|
93 |
* which consists of 16 64-bit values, can have one of 2<sup>1024</sup> values. Of those
|
|
94 |
* 2<sup>1024</sup> subsequence values, nearly all of them (2<sup>1024</sup>-2<sup>64</sup>)
|
|
95 |
* occur 2<sup>64</sup> times over the course of the entire cycle, and the other
|
|
96 |
* 2<sup>64</sup> subsequence values occur only 2<sup>64</sup>-1 times. So the ratio
|
|
97 |
* of the probability of getting one of the less common subsequence values and the
|
|
98 |
* probability of getting one of the more common subsequence values is 1-2<sup>-64</sup>.
|
|
99 |
* (Note that the set of 2<sup>64</sup> less-common subsequence values will differ from
|
|
100 |
* one instance of {@code L64X1024Random} to another, as a function of the additive
|
|
101 |
* parameter of the LCG.) The values produced by the {@code nextInt()}, {@code nextFloat()},
|
|
102 |
* and {@code nextDouble()} methods are likewise 16-equidistributed.
|
|
103 |
*
|
|
104 |
* <p>Method {@link #split} constructs and returns a new {@code L64X1024Random}
|
|
105 |
* instance that shares no mutable state with the current instance. However, with
|
|
106 |
* very high probability, the values collectively generated by the two objects
|
|
107 |
* have the same statistical properties as if the same quantity of values were
|
|
108 |
* generated by a single thread using a single {@code L64X1024Random} object.
|
|
109 |
* This is because, with high probability, distinct {@code L64X1024Random} objects
|
|
110 |
* have distinct {@code a} parameters and therefore use distinct members of the
|
|
111 |
* algorithmic family; and even if their {@code a} parameters are the same, with
|
|
112 |
* very high probability they will traverse different parts of their common state
|
|
113 |
* cycle.
|
|
114 |
*
|
|
115 |
* <p>As with {@link java.util.SplittableRandom}, instances of
|
|
116 |
* {@code L64X1024Random} are <em>not</em> thread-safe.
|
|
117 |
* They are designed to be split, not shared, across threads. For
|
|
118 |
* example, a {@link java.util.concurrent.ForkJoinTask} fork/join-style
|
|
119 |
* computation using random numbers might include a construction
|
|
120 |
* of the form {@code new Subtask(someL64X1024Random.split()).fork()}.
|
|
121 |
*
|
|
122 |
* <p>This class provides additional methods for generating random
|
|
123 |
* streams, that employ the above techniques when used in
|
|
124 |
* {@code stream.parallel()} mode.
|
|
125 |
*
|
|
126 |
* <p>Instances of {@code L64X1024Random} are not cryptographically
|
|
127 |
* secure. Consider instead using {@link java.security.SecureRandom}
|
|
128 |
* in security-sensitive applications. Additionally,
|
|
129 |
* default-constructed instances do not use a cryptographically random
|
|
130 |
* seed unless the {@linkplain System#getProperty system property}
|
|
131 |
* {@code java.util.secureRandomSeed} is set to {@code true}.
|
|
132 |
*
|
|
133 |
* @author Guy Steele
|
|
134 |
* @since 1.9
|
|
135 |
*/
|
|
136 |
public final class L64X1024Random extends AbstractSplittableRng {
|
|
137 |
|
|
138 |
/*
|
|
139 |
* Implementation Overview.
|
|
140 |
*
|
|
141 |
* The split() operation uses the current generator to choose 18 new 64-bit
|
|
142 |
* long values that are then used to initialize the parameter `a`, the
|
|
143 |
* state variable `s`, and the array `x` for a newly constructed generator.
|
|
144 |
*
|
|
145 |
* With extremely high probability, no two generators so chosen
|
|
146 |
* will have the same `a` parameter, and testing has indicated
|
|
147 |
* that the values generated by two instances of {@code L64X1024Random}
|
|
148 |
* will be (approximately) independent if have different values for `a`.
|
|
149 |
*
|
|
150 |
* The default (no-argument) constructor, in essence, uses
|
|
151 |
* "defaultGen" to generate 18 new 64-bit values for the same
|
|
152 |
* purpose. Multiple generators created in this way will certainly
|
|
153 |
* differ in their `a` parameters. The defaultGen state must be accessed
|
|
154 |
* in a thread-safe manner, so we use an AtomicLong to represent
|
|
155 |
* this state. To bootstrap the defaultGen, we start off using a
|
|
156 |
* seed based on current time unless the
|
|
157 |
* java.util.secureRandomSeed property is set. This serves as a
|
|
158 |
* slimmed-down (and insecure) variant of SecureRandom that also
|
|
159 |
* avoids stalls that may occur when using /dev/random.
|
|
160 |
*
|
|
161 |
* File organization: First static fields, then instance
|
|
162 |
* fields, then constructors, then instance methods.
|
|
163 |
*/
|
|
164 |
|
|
165 |
/* ---------------- static fields ---------------- */
|
|
166 |
|
|
167 |
/*
|
|
168 |
* The length of the array x.
|
|
169 |
*/
|
|
170 |
|
|
171 |
private static final int N = 16;
|
|
172 |
|
|
173 |
/**
|
|
174 |
* The seed generator for default constructors.
|
|
175 |
*/
|
|
176 |
private static final AtomicLong defaultGen = new AtomicLong(RngSupport.initialSeed());
|
|
177 |
|
|
178 |
/*
|
|
179 |
* The period of this generator, which is (2**1024 - 1) * 2**64.
|
|
180 |
*/
|
|
181 |
private static final BigInteger thePeriod =
|
|
182 |
BigInteger.ONE.shiftLeft(N*64).subtract(BigInteger.ONE).shiftLeft(64);
|
|
183 |
|
|
184 |
/*
|
|
185 |
* Multiplier used in the LCG portion of the algorithm, taken from
|
|
186 |
* Pierre L'Ecuyer, Tables of linear congruential generators of
|
|
187 |
* different sizes and good lattice structure, <em>Mathematics of
|
|
188 |
* Computation</em> 68, 225 (January 1999), pages 249–260,
|
|
189 |
* Table 4 (first multiplier for size 2<sup>64</sup>).
|
|
190 |
*/
|
|
191 |
|
|
192 |
private static final long m = 2862933555777941757L;
|
|
193 |
|
|
194 |
/* ---------------- instance fields ---------------- */
|
|
195 |
|
|
196 |
/**
|
|
197 |
* The parameter that is used as an additive constant for the LCG.
|
|
198 |
* Must be odd.
|
|
199 |
*/
|
|
200 |
private final long a;
|
|
201 |
|
|
202 |
/**
|
|
203 |
* The per-instance state: s for the LCG; the array x for the xorshift;
|
|
204 |
* p is the rotating pointer into the array x.
|
|
205 |
* At least one of the 16 elements of the array x must be nonzero.
|
|
206 |
*/
|
|
207 |
private long s;
|
|
208 |
private final long[] x;
|
|
209 |
private int p = N - 1;
|
|
210 |
|
|
211 |
/* ---------------- constructors ---------------- */
|
|
212 |
|
|
213 |
/**
|
|
214 |
* Basic constructor that initializes all fields from parameters.
|
|
215 |
* It then adjusts the field values if necessary to ensure that
|
|
216 |
* all constraints on the values of fields are met.
|
|
217 |
*/
|
|
218 |
public L64X1024Random(long a, long s,
|
|
219 |
long x0, long x1, long x2, long x3,
|
|
220 |
long x4, long x5, long x6, long x7,
|
|
221 |
long x8, long x9, long x10, long x11,
|
|
222 |
long x12, long x13, long x14, long x15) {
|
|
223 |
// Force a to be odd.
|
|
224 |
this.a = a | 1;
|
|
225 |
this.s = s;
|
|
226 |
this.x = new long[N];
|
|
227 |
this.x[0] = x0;
|
|
228 |
this.x[1] = x1;
|
|
229 |
this.x[2] = x2;
|
|
230 |
this.x[3] = x3;
|
|
231 |
this.x[4] = x4;
|
|
232 |
this.x[5] = x5;
|
|
233 |
this.x[6] = x6;
|
|
234 |
this.x[7] = x7;
|
|
235 |
this.x[8] = x8;
|
|
236 |
this.x[9] = x9;
|
|
237 |
this.x[10] = x10;
|
|
238 |
this.x[11] = x11;
|
|
239 |
this.x[12] = x12;
|
|
240 |
this.x[13] = x13;
|
|
241 |
this.x[14] = x14;
|
|
242 |
this.x[15] = x15;
|
|
243 |
// If x0, x1, ..., x15 are all zero (very unlikely), we must choose nonzero values.
|
|
244 |
if ((x0 | x1 | x2 | x3 | x4 | x5 | x6 | x7 | x8 | x9 | x10 | x11 | x12 | x13 | x14 | x15) == 0) {
|
|
245 |
for (int j = 0; j < N; j++) {
|
|
246 |
this.x[j] = RngSupport.mixStafford13(s += RngSupport.GOLDEN_RATIO_64);
|
|
247 |
}
|
|
248 |
}
|
|
249 |
}
|
|
250 |
|
|
251 |
/**
|
|
252 |
* Creates a new instance of {@code L64X1024Random} using the
|
|
253 |
* specified {@code long} value as the initial seed. Instances of
|
|
254 |
* {@code L64X1024Random} created with the same seed in the same
|
|
255 |
* program execution generate identical sequences of values.
|
|
256 |
*
|
|
257 |
* @param seed the initial seed
|
|
258 |
*/
|
|
259 |
public L64X1024Random(long seed) {
|
|
260 |
// Using a value with irregularly spaced 1-bits to xor the seed
|
|
261 |
// argument tends to improve "pedestrian" seeds such as 0 or
|
|
262 |
// other small integers. We may as well use SILVER_RATIO_64.
|
|
263 |
//
|
|
264 |
// The seed is hashed by mixMurmur64 to produce the `a` parameter.
|
|
265 |
// The seed is hashed by mixStafford13 to produce the initial `x[0]`,
|
|
266 |
// which will then be used to produce the first generated value.
|
|
267 |
// The other x values are filled in as if by a SplitMix PRNG with
|
|
268 |
// GOLDEN_RATIO_64 as the gamma value and Stafford13 as the mixer.
|
|
269 |
this(RngSupport.mixMurmur64(seed ^= RngSupport.SILVER_RATIO_64),
|
|
270 |
1,
|
|
271 |
RngSupport.mixStafford13(seed),
|
|
272 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
273 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
274 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
275 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
276 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
277 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
278 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
279 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
280 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
281 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
282 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
283 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
284 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
285 |
RngSupport.mixStafford13(seed += RngSupport.GOLDEN_RATIO_64),
|
|
286 |
RngSupport.mixStafford13(seed + RngSupport.GOLDEN_RATIO_64));
|
|
287 |
}
|
|
288 |
|
|
289 |
/**
|
|
290 |
* Creates a new instance of {@code L64X1024Random} that is likely to
|
|
291 |
* generate sequences of values that are statistically independent
|
|
292 |
* of those of any other instances in the current program execution,
|
|
293 |
* but may, and typically does, vary across program invocations.
|
|
294 |
*/
|
|
295 |
public L64X1024Random() {
|
|
296 |
// Using GOLDEN_RATIO_64 here gives us a good Weyl sequence of values.
|
|
297 |
this(defaultGen.getAndAdd(RngSupport.GOLDEN_RATIO_64));
|
|
298 |
}
|
|
299 |
|
|
300 |
/**
|
|
301 |
* Creates a new instance of {@code L64X1024Random} using the specified array of
|
|
302 |
* initial seed bytes. Instances of {@code L64X1024Random} created with the same
|
|
303 |
* seed array in the same program execution generate identical sequences of values.
|
|
304 |
*
|
|
305 |
* @param seed the initial seed
|
|
306 |
*/
|
|
307 |
public L64X1024Random(byte[] seed) {
|
|
308 |
// Convert the seed to 18 long values, of which the last 16 are not all zero.
|
|
309 |
long[] data = RngSupport.convertSeedBytesToLongs(seed, 18, 16);
|
|
310 |
long a = data[0], s = data[1];
|
|
311 |
// Force a to be odd.
|
|
312 |
this.a = a | 1;
|
|
313 |
this.s = s;
|
|
314 |
this.x = new long[N];
|
|
315 |
for (int j = 0; j < N; j++) {
|
|
316 |
this.x[j] = data[2+j];
|
|
317 |
}
|
|
318 |
}
|
|
319 |
|
|
320 |
/* ---------------- public methods ---------------- */
|
|
321 |
|
|
322 |
/**
|
|
323 |
* Constructs and returns a new instance of {@code L64X1024Random}
|
|
324 |
* that shares no mutable state with this instance.
|
|
325 |
* However, with very high probability, the set of values collectively
|
|
326 |
* generated by the two objects has the same statistical properties as if
|
|
327 |
* same the quantity of values were generated by a single thread using
|
|
328 |
* a single {@code L64X1024Random} object. Either or both of the two
|
|
329 |
* objects may be further split using the {@code split} method,
|
|
330 |
* and the same expected statistical properties apply to the
|
|
331 |
* entire set of generators constructed by such recursive splitting.
|
|
332 |
*
|
|
333 |
* @param source a {@code SplittableRng} instance to be used instead
|
|
334 |
* of this one as a source of pseudorandom bits used to
|
|
335 |
* initialize the state of the new ones.
|
|
336 |
* @return a new instance of {@code L64X1024Random}
|
|
337 |
*/
|
|
338 |
public L64X1024Random split(SplittableRng source) {
|
|
339 |
// Literally pick a new instance "at random".
|
|
340 |
return new L64X1024Random(source.nextLong(), source.nextLong(),
|
|
341 |
source.nextLong(), source.nextLong(),
|
|
342 |
source.nextLong(), source.nextLong(),
|
|
343 |
source.nextLong(), source.nextLong(),
|
|
344 |
source.nextLong(), source.nextLong(),
|
|
345 |
source.nextLong(), source.nextLong(),
|
|
346 |
source.nextLong(), source.nextLong(),
|
|
347 |
source.nextLong(), source.nextLong(),
|
|
348 |
source.nextLong(), source.nextLong());
|
|
349 |
}
|
|
350 |
|
|
351 |
/**
|
|
352 |
* Returns a pseudorandom {@code long} value.
|
|
353 |
*
|
|
354 |
* @return a pseudorandom {@code long} value
|
|
355 |
*/
|
|
356 |
|
|
357 |
public long nextLong() {
|
|
358 |
// First part of xoroshiro1024: fetch array data
|
|
359 |
final int q = p;
|
|
360 |
final long s0 = x[p = (p + 1) & (N - 1)];
|
|
361 |
long s15 = x[q];
|
|
362 |
|
|
363 |
final long z = s + s0;
|
|
364 |
s = m * s + a; // LCG
|
|
365 |
|
|
366 |
// Second part of xoroshiro1024: update array data
|
|
367 |
s15 ^= s0;
|
|
368 |
x[q] = Long.rotateLeft(s0, 25) ^ s15 ^ (s15 << 27);
|
|
369 |
x[p] = Long.rotateLeft(s15, 36);
|
|
370 |
|
|
371 |
return z;
|
|
372 |
}
|
|
373 |
|
|
374 |
public BigInteger period() { return thePeriod; }
|
|
375 |
}
|